Servicing the Server

This chapter contains the following topics:

Maintaining the Server

To keep server maintenance simple, the server contains a set of components that are field-replaceable. However, not all components are field-replaceable.

For components that are not a field-replaceable unit (FRU), contact your Cisco account team to inquire about replacements.

Use the following tasks for field-service:

Removing Top Cover

Before you begin

For this procedure, gather a #2 Phillips screwdriver.

Procedure


Step 1

Using a Phillips screwdriver, loosen the M3 thumbscrew on the top cover.

Step 2

Slide the cover to back and detach the cover from the small knob screws along the sides and back of the server.

Step 3

Lift the top cover straight up to remove it from the server, then set the cover aside.

Figure 1. Removing the Top Cover

1

Top Rear Cover


Service Considerations

Be aware of the following when servicing the server:


Important


Watch your hands and fingers whenever you handle the server, sleds, and components! Narrow vertical or horizontal spaces in situations like, but not limited to, moving the server into or out of the shipping container or equipment rack can cause pinch hazards for your hands and fingers.



Warning


The fully configured server weighs approximately 265 lbs. (120 kg)! Never attempt to lift the server by yourself! Instead, use a lift, scissors jack, or some other device to lift and bear the weight of the server while you are installing or servicing it. If you must lift or handle the server without mechanical assistance, use two or more people to lift and handle server. Always use safe lifting practices when lifting or moving the server.



Note


When handling chassis components, wear an ESD strap and handle components, modules, and sled by the edges, or by designated handles when present.


The server is designed as a modular system with subsystems for compute, cooling, and so on. Subsystems are generally arranged in trays, which are field replaceable units (FRUs). However, some components in a subsystem or tray are not individually field-replaceable. The following main components are not field replaceable.

  • Mainboard

  • Power distribution board

  • Mid-plane board

  • HIB host interface Board

  • FIO board

  • Fan extension board

  • Fan interposer board

Hot Swapping the Fan Module

The server has multiple fan modules for proper cooling. Fans direct airflow front-to-back, and each fan is a hot swappable FRU. Fan replacement is a tool less procedure.

Use the following task to remove and replace a fan module.

Procedure


Step 1

Remove the hot swap fan module:

  1. Press the latch to disengage the fan module from the cage

  2. Pull the handle to detach the fan module from the server.

    Figure 2. Removing the Fan Module

Step 2

Replace the hot swap fan module:

  1. Install the hot swap fan module:

  2. Push the new fan module into the cage until it locks into place.

    Figure 3. Replacing the Fan Module

Removing and Replacing 12V Redundant PSU


Caution


Replace all cut cable ties during removal and replacement procedure. Failure to do so results in improper airflow that can lead to thermal damage.



Warning


To reduce the risk of injury from electric shock, remove all power cords to disconnect power from the system completely.


The server's 12V PSUs are toolless FRUs. Use this task to replace a 12V PSU.

Procedure


Step 1

Remove the 12V redundant power supply unit:

  1. Press the latch to disengage the power supply from the CPU tray.

  2. Pull the handle to detach the power supply while still pressing the latch.

    Figure 4. Removing the 12V Redundant PSU

Step 2

Replace the 12V redundant power supply unit:

  1. Install the 12V redundant power supply unit:

  2. Push the new power supply into the CPU tray until it locks into place.

    Figure 5. Replacing the 12V Redundant PSU

Removing and Replacing the OCP Module

The OCP Module is part of the CPU Tray above the DC-SCM Module.

OCP Module

PID

UCS C845A M8 One OCP 3.0 SFF PCIe Gen5 x8 NIC (CPU0) with two 10GbE RJ-45 Ethernet

CAI-O-ID10GC

Use the following task to replace the OCP Module.

Before you begin

For this procedure, gather a #2 Phillips screwdriver.

Procedure


Step 1

Remove the OCP Module.

  1. Using a Phillips screwdriver, loosen the M3 thumbscrew.

  2. Grasp the OCP module and slide it out of the CPU Tray.

    Figure 6. Removing the OCP Module

Step 2

Install the OCP Module.

  1. Align the OCP Module with its slot on the CPU Tray.

  2. Holding the module level, slide it into the CPU Tray.

  3. When the module is flush with the CPU Tray, use the Phillips screwdriver to tighten the thumbscrews.

Figure 7. Installing the OCP Module

Removing and Replacing the E1.S NVMe Drive

The server supports E1.S NVMe SSDs in the front-loading drive bays.

Before you begin

For this procedure, gather a #2 Phillips screwdriver.

Procedure


Step 1

Remove the E1.S NVMe SSD carrier assembly from the chassis:

  1. Push the drive carrier button to disengage the drive carrier bezel.

  2. Pull the drive carrier bezel to detach the carrier assembly from the chassis.

  3. Remove the carrier assembly from the chassis.

    Figure 8. Removing the NVMe/SSD Carrier Assembly

Step 2

Remove the E1.S NVMe SSD drive from the carrier:

  1. Using a Phillips screwdriver, remove the two (2) screws.

  2. Remove the E1.S NVMe SSD drive from the carrier.

    Figure 9. Removing the E1.S NVMe SSD from the Carrier Assembly

Step 3

Install E1.S NVMe SSD drive to the carrier:

  1. Place the NVMe/SAS SSD drive into the chassis.

  2. Using a Phillips screwdriver, insert and tighten the two (2) screws to secure the NVMe/SAS SSD drive to the carrier.

    Figure 10. Replacing the NVMe/SSD in the Carrier Assembly

Step 4

Install E1.S NVMe SSD carrier assembly to the chassis:

  1. Place the E1.S NVMe SSD carrier assembly into the chassis.

  2. Push the carrier assembly into the chassis until it latches into place.

    Figure 11. Replacing the NVMe/SSD Carrier Assembly

Replacing the M2 Internal Storage

The server has an internal M.2 module that sits on the air baffle in the middle of the module. The module consists of two individual M.2 SSDs that are installed vertically on the M.2 carrier and are attached by a securing screw at one end of the SSD. The other end of each SSD seats into a socket on the carrier.

  • Drive 0 is the SSD on the top of the carrier

  • Drive 1 is on the bottom of the carrier

To replace the M.2 module and individual SSDs, use the following tasks.

Removing Internal M.2 Storage

The server has an internal M.2 module that sits on the air baffle in the middle of the module. Use the following procedure to remove the internal M.2 Module.

Before you begin

If you have not already disconnected the server from facility power, do so now.

You will find it helpful to gather a #2 Phillips and an M3 Phillips screwdriver.

Procedure


Step 1

Remove the top cover.

Removing Top Cover.

Step 2

Remove the top support bracket.

  1. Locate the two M3 thumbscrews, and use a #2 Phillips screwdriver to remove them.

  2. Slide the support bracket, then lift it up to release it from the chassis.

Step 3

Remove the cage.

  1. Slide it so that it disengages from the T pins.

  2. Lift it off of the server.

Step 4

Remove the fan cage.

  1. Unplug the fan cables from the HSC board.

  2. Simultaneously raise each lever of the fan cage, then lift it out of the chassis.

Step 5

Remove the cables from the HSC Board.

  1. Locate and disconnect the cable ends shown below.

    1

    SW Board cables, two

    2

    PDB Board cables, three

    3

    BF3 NIC/ConnectX-7 cable, 1

    4

    GPU Power Cables

  2. Pull the cables aside so that they do not create an obstruction.

Step 6

Remove the HSC Module.

  1. Locate the two M3 thumbscrews, and use a #2 Phillips screwdriver to remove them.

  2. Remove the HSC Module by sliding it, then lifting it out of the chassis.

Step 7

Remove the M.2 Module.

  1. Locate the M.2 module.

  2. Remove the following cable.

  3. Remove the following cable.

  4. Disconnect the screws and clear the cables so that there are no obstructions, then remove the M.2 module.

Step 8

When the M.2 module is removed from the chassis, remove one or both SSDs.

  1. Using an M3 Phillips screwdriver, remove the securing screw (circled in the left image of the following illustration).

  2. Simultaneously, slightly lift the end of the SSD where you removed the screw and slide the other end (connector side) of the SSD out of the socket.

    Caution

     

    Perform this step carefully to avoid damaging the SSD or carrier!

    1

    Drive Slot 0

    2

    Drive Slot 1


What to do next

When you are ready, insert one or more M.2 SSDs and replace the M.2 Module in the chassis.

Go to Installing Internal M.2 Storage.

Installing Internal M.2 Storage

Use the following procedure to install the internal M.2 SSDs and place the M.2 module in the chassis.

Before you begin

If you have not already disconnected the server from facility power, do so now.

You will find it helpful to gather a #2 Phillips and an M3 Phillips screwdriver.

Procedure


Step 1

Install the SSDs onto the M.2 module.

  1. Position the SSD in either the top (Drive 0) or bottom (Drive 1) slot of the carrier as shown in the table below the following illustration.

  2. Gently slide the golden fingers of the SSD into the connector, then gently press the other end down until the SSD is flush with the carrier.

    When installed correctly, the SSD lies flat on the carrier and the notch on the SSD is aligned with the screwhole for the securing screw.

  3. Using an M3 Phillips screwdriver, install the securing screw to fasten the SSD to the carrier as shown by a circle in the left image of the following illustration.

    1

    Drive Slot 0

    2

    Drive Slot 1

Step 2

Install the M.2 Module and connect the required cables.

  1. Position the M.2 module.

  2. Connect the following cable.

  3. Connect the following cable.

  4. Using the #2 Phillips screwdriver, tighten the screws to secure the M.2 module to the chassis.

Step 3

Install the HSC Module.

  1. Align the HSC Module and slide it into the chassis.

  2. Install the two M3 thumbscrews using a #2 Phillips screwdriver to secure the module.

Step 4

Connect the cables to the HSC Board.

  1. Route the cables to their respective connection points.

  2. Connect the cable ends to the HSC Board.

Step 5

Install the fan cage.

  1. Lower the fan cage into the chassis and press down until the levers lock into place.

  2. Plug the fan cables into the HSC board.

Step 6

Install the cage.

  1. Lower the cage onto the server.

  2. Slide the cage so that it engages with the T pins.

Step 7

Install the top support bracket.

  1. Align the support bracket and slide it into the chassis.

  2. Install the two M3 thumbscrews and tighten them using a #2 Phillips screwdriver.

Step 8

Install the top cover.

Removing Top Cover.


Replacing the DC-SCM Module

The DC-SCM Module is a field-serviceable part.

Although the following are similar, there are some differences due to hardware information that changes depending on the workflow:

Removing and Replacing the DC-SCM Module

The DC-SCM Module is part of the CPU Tray below the OCP Module. The module is a FRU that is offered in the following options depending on the type of CPU and GPU installed in the server.

DC-SCM Module

PID

UCS C845A M8 DC-SCM Card Module supporting NVIDIA GPU, 5th Gen AMD EPYC CPU

CAI-845A-DCSCM(=)

Use the following task to replace the DC-SCM Module.

Before you begin

For this procedure, gather a #2 Phillips screwdriver.

Procedure


Step 1

Remove the DC-SCM Module.

  1. Using a Phillips screwdriver, loosen the M3 thumbscrew.

  2. Grasp the DC-SCM module and slide it out of the CPU Tray.

    Figure 12. Removing the DC-SCM Module

Step 2

Install the DC-SCM Module.

  1. Align the DC-SCM Module with its slot on the CPU Tray.

  2. Holding the module level, slide it into the CPU Tray.

  3. When the module is flush with the CPU Tray, use the Phillips screwdriver to tighten the thumbscrews.


DC-SCM Swap-In Considerations

When replacing a DC-SCM module with a new one, for example, in the unlikely event of a failure, be aware of the following :

  • The new DC-SCM module has different hardware information than the original module. Therefore, you must update the chassis labeling by attaching the new labels that accompany the new DC-SCM.

  • Two new labels will be provided with the new DC-SCM. Each label contains the same information, the MAC address of the new DC-SCM.

  • Make sure to use the new MAC address and affix it to the chassis in both of the appropriate locations:

    • The asset pull tag, which is accessible from the front of the chassis.

    • The chassis label which is located in the center for the chassis top cover.

    • Affixing the label for the new DC-SCM's MAC address requires partially removing the chassis from the rack to gain access to the location of the existing label. You should not need to completely de-rack the chassis if you are using the supplied rail kit. Instead, to access the chassis label, you just need to slide the chassis out of the equipment rack on the supplied rail kit.

  • When referring to the DC-SCM MAC address or scanning barcodes, be aware of the following:

    • Use only the new MAC address. Do not use the old MAC address.

    • The chassis label contains a QR code that points to online product information through Cisco's easy-to-use Documentation Portal.

      When scanning the QR code, be aware that the existing chassis PID (product identifier), VID (version identifier), and SN (serial number) shown on the Documentation Portal are still valid after swapping in the new DC-SCM. However, the MAC address shown on the Documentation Portal will no longer be valid after swapping in the new DC-SCM because that information still shows the old MAC address.

To swap in a new DC-SCM, go to Swapping In a New DC-SCM Module.

Swapping In a New DC-SCM Module

When you swap out an old DC-SCM and swap in a new one, the new DC-SCM has different information than the DC-SCM that you are swapping out, so it is critical that you make the appropriate changes for record keeping purposes. For more information, see DC-SCM Swap-In Considerations.

Before you begin

Use this task to swap out an old DC-SCM and swap in a new one and affix the labels for the new DC-SCM to the chassis.


Caution


Adding the labels for the new DC-SCM requires partially removing the chassis from the rack. Be aware that to complete this procedure, the server might experience a service interruption. As a best practice you should schedule this procedure during a regularly scheduled maintenance window to reduce the effect of downtime.


Gather the following before attempting this procedure:

  • The new DC-SCM package, which contains the new module, plus two new labels.

    • One larger label has the DC-SCM MAC address. This label should be affixed to the top of the chassis.

    • One smaller label also has the DC-SCM MAC address. This label should be affixed to the chassis asset tag next to the serial number.

  • A #2 Phillips screwdriver

  • A permanent marker

Procedure


Step 1

Prepare to access the server.

  1. Disconnect any cables that might obstruct sliding the chassis out of the rack.

  2. Using a #2 Phillips screwdriver, remove the screws that secure the chassis to the rack.

  3. Gather all the screws and put them somewhere safe,

    You will use these screws to re-install the chassis into the rack.

Step 2

Slide the chassis partially out of the rack so that you can access the label on the top right corner of the chassis.

When properly installed, the chassis should be easily slide out because of the slide rail kit.

If you encounter any resistance, disengage the locking mechanism on both slide rails.

Step 3

Swap in the new DC-SCM.

  1. Using the Phillips screwdriver, loosen the captive M3 screws on the old DC-SCM.

  2. Remove the old DC-SCM and install the new one.

  3. Using the Phillips screwdriver, tighten the captive M3 screws to snug.

Step 4

Attach the new chassis label.

  1. Locate the chassis label on the top of the server chassis.

  2. On the chassis label, use a permanent marker to obscure the old DC-SCM's MAC address.

  3. Attach the new DC-SCM's MAC address label, which is the larger label, below the existing label.

    Caution

     

    Do not paste the new label over the old label. The old label still has usable product information. If you attach the new label to the wrong place, you can obscure product information that is still valid.

Step 5

Attach the new asset tag label, which is the smaller label.

  1. Locate the asset tag at the top left corner of the chassis front panel.

    The asset tag is a heavy plastic pull tab that slides into and out of a vertical slot on the chassis.

  2. Slide the asset tag out and put the new DC-SCM's MAC address label directly over the old MAC address label next to the serial number.

  3. Slide the asset tag back into the chassis, making sure that the label does not bind, get damaged, or peel off while you insert the asset tag.

Step 6

Reinstall the chassis.

  1. Slide the chassis back into place in the rack.

  2. Reinsert the screws through the chassis mounting brackets and into the screwholes in the equipment rack.

    For reference, see Step 7 through Step 9 of Installing the Server.
  3. Reconnect any cables that you disconnected.


Removing and Replacing the GPU

Before you begin

For this procedure, gather a #2 Phillips screwdriver.

Procedure


Step 1

Remove the rear top cover from the chassis. Removing Top Cover

  1. Remove the two screws.

  2. Detach the rear top cover from the chassis.

    Figure 13. Removing the cover from the chassis

Step 2

Remove the GPU:

  1. Detach the GPU from the PCIe slot.

    Figure 14. Removing GPU

Step 3

Replacing the GPU

  1. Place the GPU into the PCIe slot and push it down until it latches into place.

    Figure 15. Replacing the GPU

Step 4

Install the rear top cover:

  1. Place the rear top cover onto the chassis.

  2. Secure the rear top cover to the chassis using two screws.

    Figure 16. Replacing Rear Top Cover